Projects to sequence vertebrate genomes are proceeding more rapidly than was imagined a few years ago. New computational tools for comparative genome analysis at the nucleotide level are needed to more effectively identify functional but non-protein-coding segments and to dissect mammalian evolution. The Penn State group has excelled at developing such tools. Our PipMaker and MultiPipMaker Web servers set the standard for alignment of user specified genomic sequences, and our Blastz program was chosen to produce high-sensitivity alignments for the Mouse Genome Analysis Consortium. We were also a significant source of biological and statistical expertise within the Consortium, particularly with respect to functional non-coding segments and evolution. We will raise comparative genome studies to a higher level by developing software that accurately identifies the full spectrum of mutational events. Current multiple alignment procedures use one sequence as the reference and hence give an asymmetric and incomplete view of sequence relationships. Our new Generalized Multiple Alignments will provide symmetric and complete views of the alignments, accurately identify kilobase-scale insertions and deletions, and permit any of the species to be used as a reference in subsequent analysis. The results obtained by our new alignment programs will be analyzed using new statistical procedures to more accurately predict the locations of elements that regulate gene transcription, and to measure variation and co-variation of mutational rates along the genome. These computational studies will guide experimental work to confirm regulatory sites and to identify the biological mechanisms that underlie rate variations in neutral evolution. An integral part of this effort will be our continued collaborations with the NISC Comparative Sequencing Program, with the group headed by David Haussler and Jim Kent at the University of California at Santa Cruz, and with the Comparative Chloroplast Genomics Project.